8/12/2019 2949313 mercado

1/17

Anthropological Studies of Cities

Author(s): Richard E. BlantonReviewed work(s):Source: Annual Review of Anthropology, Vol. 5 (1976), pp. 249-264Published by: Annual ReviewsStable URL: http://www.jstor.org/stable/2949313.

Accessed: 01/02/2012 13:00

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at.http://www.jstor.org/page/info/about/policies/terms.jsp

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of

content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new formsof scholarship. For more information about JSTOR, please contact support@jstor.org.

Annual Reviewsis collaborating with JSTOR to digitize, preserve and extend access toAnnual Review of

Anthropology.

http://www.jstor.org

http://www.jstor.org/action/showPublisher?publisherCode=annrevshttp://www.jstor.org/stable/2949313?origin=JSTOR-pdfhttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/stable/2949313?origin=JSTOR-pdfhttp://www.jstor.org/action/showPublisher?publisherCode=annrevs

8/12/2019 2949313 mercado

2/17

Ann.

Rev.

AnthropoL

1976. 5:249-64

Copyright ?

1976

by

Annual Reviews Inc. All

rights

reserved

ANTHROPOLOGICAL STUDIES

*

9579

OF

CITIES

Richard E. Blanton

Departmentof Sociology and Anthropology,Purdue

University,

West Lafayette, Indiana47907

INTRODUCTION

In this paper I will review and expand upon a theoretical and methodological

approach

to

the study of cities that, in my view, has been unduly neglected by

anthropologists interested

in

cities. This approach, usually referred to as

re-

gional analysis, is based on the central place theory of geography.

I

believe

that

we need to

expand our theoretical horizons because traditional theories

of cities

utilized by anthropologists are both misleading and limited in scope. In order

to

demonstrate the value

of

a regional approach to studies of cities,

I

will

first

critically (and very briefly) review traditional theories

of

cities,

then offer

a

regionally based definition of cities and explore some of the implications

of the

regional approach for explaining the cross-cultural variability

and

dynamic

features of

cities and systems

of

cities.

Perhaps the most prominent feature of traditional theories

is that

they regard

cities as the sources

of

social

change-they

are

a kind of

special environment,

almost

with a

life of their own.

These city-centric

theories

have been based

on

Durkheim's

The Division of Labor in Society (who,

in

turn,

was influenced

by

Spencer's Principles of Sociology),

on the

sociology

of

George Simmel,

and the

philosophy

of

Oswald Spengler. Martindale (32)

has

written a

convenient review

of the development of these theories, which he refers to as socio-psy-

chological

theories

of cities.

Hauser

(19),

for

example, following

Durkheim

and

Spencer,

writes that

.

. .

aggregative living

has

produced

in the

social

realm

a

major transformation

the

equivalent

of

genetic

mutation

in

the

biological

realm. Park,

in

the

classic statement

of the

Chicago

School

of

urban

sociology (38), described

the

city as a psychophysical mechanism ;

Wirth

(59),

in

the same volume, writes ofthe city that

its . . .

growth

is

so

rapid

and its

energy

so

great

that

it

changes

its

complexion

almost

daily, and,

with

it,

the

character

of

mankind

itself'

(see also 60).

As

proposed by

Redfield &

Singer

(42),

the

city,

whether the

center of orthogenetic change ( .

.

.

carrying

forward into systematic and reflective dimensions an old culture ) or hetero-

genetic change

( .

. .

creating of original

modes

of thought

that have

authority

beyond

or

in

conflict

with old

cultures and civilizations ),

is

. . .

a place

in

which culture

change takes place (42, p. 58).

249

8/12/2019 2949313 mercado

3/17

250 BLANTON

Robert Redfield's famous Folk-Urban Continuum (41)

was developed

under the influence of the Chicago School. This proved

to be the focus

of

anthropological theorizing about

cities for years, until gradually

the work of a

number of people (too numerous to review completely here) demonstrated that

not only could the Folk-Urban Continuum not be attributed

the status

of

theory, but that it could not even stand as an adequate empirical

generalization

(cf Hauser 20; Lewis 28, 29).

In spite of the serious criticisms which have been leveled at the

Simmel-Wirth-

Redfield trio, there is no shortage of socio-psychological theories

of cities in

the more recent literature in anthropology and its sister disciplines.

Friedmann

(15) writes that cities are

. . . an active force in the ongoing processes of social

transformation, and, according to Epstein (13), most urban

anthropologists

in

Africa assume that . . . towns inevitably act as instruments of social trans-

formation (see also Jacobs 22 and Moore 37). Although many

criticisms could

be made of these traditional ways of looking at cities, the major

objections,

as

I

see

them,

are:

(a)

If

urbanism

is regarded as a way of life, how can one account

for the cross-cultural and diachronic variability

in

cities and

in the

ways

of life

within

cities?

The

attempt

by Redfield & Singer (42) to differentiate between

cities which are the centers of orthogenetic or heterogenetic

change by

no

means does justice to the wide variety

in

kinds and

functions of cities. (b)

As

stated

by

Lewis

(28, p. 432),

the

. . .

folk-urban

conceptualization

of social

change

focuses attention

primarily

on

the

city

as the source

of

change,

to the

exclusion or neglect of other factors of an internal or external nature. The folk-

urban continuum fails to aid

our understanding of the rates of social change,

or

when and where

change

is likely or not likely to occur.

In

the remainder of this paper

I will attempt to demonstrate that a regional

theory of cities,

in

contrast

with socio-psychological theories,

is explicitly

oriented to

provide explanations

for the cross-cultural differences

in

cities,

while avoiding the fallacy of overemphasizing the role of cities

in

social change.

CITIES DEFINED

I

will assiduously avoid the use of the terms urban and urbanism,

for although

both these terms have

something

to do with

cities,

no two researchers

agree

on

what the association is. Similarly, much

ink

has been spilled

by scholars

over

what cities are and what

they

are not, without, to date, a satisfactory

resolution

of

the definitional

problem.

Instead of reviewing this

bulky and frustrating

literature

here,

I

refer the

reader to two useful summaries by

Paul

Wheatley (57,

pp. 371-99; 58).

Perhaps

the

greatest obstacle

in the definition of cities- as

a special type

of

human community has been the establishment of a set of indispensable criteria

which can be

applied

cross-culturally.

In the

most widely

read

work on

pre-

industrial cities

(Sjoberg 47),

for

example, it is argued that

the presence

of a

literate elite is the

single best criterion to distinguish cities

from other types

of

8/12/2019 2949313 mercado

4/17

STUDIES OF CITIES 251

early settlements, clearly a cross-cultural error

of

no

small magnitude. Wheatley

(56) suggests that these definitional problems could

be resolved

by

the use of his

concept of ethnocity. Here attention is paid

to . . . nodes of

concentration

of people and shelters

in

the continuum

of

population

distribution over the face

of the earth. Such of these as attain

a

certain size

and perform appropriate

functions are designated

by the terms appositely translated

in

English

as

'city'

or

'town'. ..

.

These nodes

are induced by forces operating within,

and

sometimes

peculiar to,

the

specific culture (56, pp.

166-67).

This approach

is

desirable

because it avoids the necessity for discovering indispensable

criteria,

and in-

stead focuses on the hierarchy of central places

in a society.

Such a functional definition of cities emphasizes the disposition

in

space

of

what might be called

central institutions-institutions that mediate between

specialized subsystems within a society. In order to optimally service a popu-

lation, these central institutions

are not likely to be randomly dispersed

over

the

landscape. Instead,

they will tend to occur clustered

in places

that become

the

central places of the society, or, in other words, the

cities

and towns. Cities

(and

towns) then are a product

of the process called segregation,

one of

the

two

basic

cultural evolutionary processes

described by Flannery (14). Segregation, ac-

cording

to

Flannery,

has to do with

.

. .

the amount of internal differentiation

and

specialization

of

subsystems

.

. .

(The

other

basic evolutionary process

described by Flannery,

centralization, refers to the

.

. .

degree of linkage

between

the

various sub-systems

and the highest-order

controls in the society

',

and is less relevant to the discussion at hand). In the present context,

segregation refers

to

the

extent to which households or groups

of households are

independent-in more evolved systems they are less

independent, necessitating

various kinds of linkages in the society between specialized

subsystems. Link-

ages between specialized

subsystems may take the form of transactions

in

the

context of

central institutions. Chieftainships, governments,

and markets are

examples

of such

central

institutions.

The

general

description

of

central institu-

tions is that

they

link

specialized subsystems by

the conversion of inputs to

outputs via a set of transactions

(cf Meier 34). The kinds of inputs and outputs

vary-markets involve mostly material inputs and outputs, while governments

more often convert information inputs to outputs

in the form of directives.

For

our

purposes here,

the most salient characteristics of central institutions

are

that

they require

energy to function, and that the transactions take time.

Energy is supplied by subsystems of producers,

who must work more than

would

be

necessary

in the absence of such institutions. The fact that there

is

always

a

finite amount of energy in the environment

of any society, and

that

producers can be pushed or otherwise encouraged

to produce only so much

surplus, means

that

central institutions always have

a maximum size and

are

always

limited

to

a finite number of

transactions per

unit of time. These energy-

related limitations can be circumvented by technological or organizational

changes

that

bring

energy savings or increase the society's ability

to

capture

energy

from its

environment,

but new

limits

will

be reached

if

central

institutions

continue to

grow.

8/12/2019 2949313 mercado

5/17

8/12/2019 2949313 mercado

6/17

STUDIES

OF CITIES 253

population

size of

central places,

as geographers

have shown (e.g.

Gun-

awardena 16, Stafford

53)

will tend to vary depending

on the number

of central

place functions

present. In a society

with a hierarchy of central places, cities

are

those communities in the highest range of the hierarchy, while towns are those

communities

occupying the middle and

lower ranges of the hierarchy.

Where the

researcher draws

the boundary between what are called

cities and what are

called

towns is

always arbitrary and

will

vary

from society to society.

The regional approach

to the definition

of cities is not peculiar to Wheatley,

myself, or others

who have borrowed

from central place studies in geography.

Wolf

(61),

for

example, has written that the city is

. . .

a

settlement

in which a

combination of functions are exercized, and which becomes

useful because

in

time greater efficiency

is obtained by

having these functions concentrated

in

one

site. And, according to Arensberg (3), the city . . . is a permanently massed,

large concentration of people in a community

having nodal

function or func-

tions, somehow

providing

for

the lacing

together (not necessarily

the sub-

ordination)

of some hinterland of the

other, perhaps lesser

communities

of a

society. My

only complaint

with Arensberg's definition

is

that it

includes

the

caveat that a city must have a large

and massed concentration

of

people.

This begs the question of how large is

large enough, and how

massed is massed

enough

for a

community

to be

called

a city.

The

advantage

of the

functional

definition of cities

I

presented above

is that any community

that is a

central

place

is a city or town

(depending on its place in the central place

hierarchy

of the

society), irrespective of its form or population size. Central places in societies

smaller

and

less

segregated than our

own may lack large

or massed

central

places,

but

they

still have what Arensberg

refers to as

nodal

functions,

and so

should be referred to as

cities and

towns.

I

should point

out in concluding this

section that while I have borrowed from

central

place

studies

in

geography

in

developing this definition

of cities,

I

have

tried to word it to avoid the overemphasis

on the movement

of goods through

systems, or,

in other

words,

the economic functions of central places that,

in

my

opinion,

characterize this field (cf

Haggett 17). Part of Christaller's original

work on the

geometry

of

central place systems

had to

do with what he

called

the

administrative

or K=7

principle

(10), which

I

will

discuss

in

more detail

below. Unfortunately, more recent

work

by geographers

has

dealt

almost ex-

clusively

with the

marketing

and

transport principles,

both concerned

with

tertiary

economic functions. The kinds of transactions

I

am

referring

to

in

the context of central institutions involve

the movement of goods and/or

infor-

mation. The overemphasis

on economic functions of central

places has led,

in

my view, to some

confusion in the literature. Coe (11), for

example, wrote that

interior Cambodian

and lowland Classic

Maya civilizations had no true

cities. His

argument

is that

since

both geographic areas lacked

environmental

diversity it is not likely that there was regional economic specialization and

trade, and therefore both areas lacked

cities. This is not

only overly environ-

mentally deterministic

(since he has almost no direct archaeological

evidence

for

the lack of

specialization

and trade),

but it also overlooks the fact that there were

8/12/2019 2949313 mercado

7/17

254

BLANTON

obviously central places,

manifested

as large groupings

of

civic-ceremonial

buildings and

palaces. Even if it were

true that

these were not economic

centers,

they were clearly centers

of information

processing and so should be referred to

as cities as defined above. Donna Taylor (54) has demonstrated for a group of

East African middle

range

hierarchical societies a . . .

close

positive re-

lationship between

settlement

hierarchy and

political-jurisdictional

hierarchy

and

that

. . .

in hierarchical

societies the relative

size of the

centers of

higher rank is

closely

related to the extent of

centralization in

decision-pro-

cessing (54, p. 13).

Interestingly, this

set of East African

chiefdoms and simple

states, unlike those

portrayed in the generalized

descriptions

of societies

of

this

type

by

Service (46) and Sahlins

(44), .

. .

were not typically redistributive

societies.

.

.

.

Local

groups tended

to be largely

self-sufficient in material ...

needs, and the organized collection and redistribution of regional production

specialties

is

described for only one

of the most centrally

organized

societies

examined

(54, p. 81).

From the same

perspective I must

add a criticism to

Wheatley's

contention

that

the

early

cult

centers

(as

he calls

them)

that

evolved

in

regions of

primary state

formation were centers of

redistribution

(57,

p. 389). The

presence or

absence of redistribution

in ancient

central places must

always

remain

hypothetical until

demonstrated by

direct archaeological

evidence.

In

the next section I

will explore

some of the

interrelationships

between

information

processing

central place

hierarchies and

economic central place

hierarchies. This point of view, I argue, will contribute to an understanding of

the nature and

dynamic properties of

systems of

cities and towns.

THE

LOCATIONAL

PATTERNS

Anthropologists interested

in

the application of

central place

theory to problems

in

peasant

marketing systems have found

Christaller's (10) original

formulation

most applicable

(cf

Crissman 12, Hodder &

Hassall 21, Johnson 24,

Skinner 48,

and

Smith

52).

According

to

some

geographers (cf

5), the transformations

of the

Christallerian model

proposed by Losch

(30)

may prove

to

be

a

more

accurate

representation of reality for

modern

industrialized

societies, but there

is

no

concensus

on

this

point.

For

these

reasons,

and because

Losch

was

relatively

less

interested

in

administrative

location,

I will

also depend

here

primarily

on

Christaller's

formulation.

Rather

than

repeat

the

assumptions

and

geometric

properties of the

Christallerian model, I

refer the reader to a recent

synthesis

by

Smith

(5 1).

The

major difference

between Christaller's two

economically

relevant

lo-

cational

principles (the

K=3

marketing principle,

and

the

K=4

transport

principle) on the

one hand, and the

K=7

administrative principle

on

the other

hand, has to do with the extent to which competition can occur between centers

of

the same level

in

the

hierarchy. (The

K

numbers

refer to the ratio

of low-order

places to

high-order

places, a matter not

relevant

to

this

discussion.

I

use the

K

designations

only

as a

convenient means

of

referring

to

the

different

locational

patterns.)

In

the case of the two

economic

principles,

centers

of

a lower level

in

8/12/2019 2949313 mercado

8/17

STUDIES OF CITIES 255

the settlement

hierarchy are nested midway in space between centers

of

the

next highest rank.

In the K=3 case, such centers are

located between three

higher-level centers, while in the case of the K=4

pattern, such centers are

located midway between two higher level centers, on the road between them.

Centers of the

same level

in

hierarchy, therefore, compete

to service

lower-level

centers, particularly

near the edges of their maximum spheres of influence. As a

consequence, all

consumers tend to be served equally,

without monopolistic

pricing. As

Christaller correctly perceived, the locations

of administrative cen-

ters should differ

from the K=3 or K=4 patterns, since they do not compete for

customers ; instead, they should have discrete, tightly

bounded ranges, with

the center conveniently situated near the center of the

servicing region. As

Smith

(52, p. 98)

notes,

the

K=7 principle

.

.

.

is

admirably suited to carving up

administrative districts, since each higher-order center controls its dependent

centers

and

hinterland

exclusively. Although I will not deal explicitly here with

religious-administrative hierarchies (which are usually

inextricably woven with

political administration), in those cases where they are

discrete, autonomous

institutions (such as

in Medieval Europe), they should exhibit the K=7 pattern,

since each parish

is a discrete unit, and competition is not likely to occur.

IMPLICATIONS

OF THE LOCATIONAL

PATTERNS FOR

UNDERSTANDING VARIABILITY

IN

CENTRAL

PLACE

HIERARCHIES

In

the following sections, I will outline the basic patterns of

articulation

between

economic

central

institutions and decision-making central institutions,

and ex-

plore the

implications of these patterns for understanding some

of the cross-

cultural

variability

in

sizes and functions of central

places,

and certain

of

their

dynamic

features.

I

do this

in

three sections.

The first two

sections,

which

deal

with

primate

centers and disembedded capitals, pertain

to the

interaction

of economic

and

decision-making central institutions

at

the

upper

levels

of

central

place

hierarchies,

while

the last section deals

with

the

patterns

of

articulation of these central institutions on the lower levels.

Primate Centers

Purely from

theoretical considerations, one should not expect locational

iso-

morphism between

administrative and economic central place hierarchies,

since

the two locational

patterns are

incompatible (cf Skinner 48, p. 31).

There are

two

major exceptions to

this general rule.

REDISTRIBUTIVE SYSTEMS

In

redistributive

economic

systems,

in which ex-

change

is

entirely

administered,

the

centers

of

redistribution

are also

centers

of

decision-making.

PRIMATE

SYSTEMS In a marketing central place

hierarchy

in which the

range

of the

highest

order

functions is equivalent to the maximum

extent

of the

society

as

a

whole, only one

highest ranking center will exist, since only that center

will

8/12/2019 2949313 mercado

9/17

256 BLANTON

be able to offer the entire range

of products n the system. This centercan also be

the highest rankingadministrative enter, since the maximum

range of these

functions is also equivalent to

the maximumextent of the whole

society. The

combinationof the widest possiblerangeof economic and administrativeunc-

tions in one place can produce a primate distribution in

the settlement

hierarchy, .e. a situation n which

the major entralplace is exceptionally arge

relativeto other centers (Berry

4). A largecenter of this type can

actuallyretard

the growth of lower-ranking

enters, thereby intensifying its

regional domi-

nance,

for

the following reasons:

(a)

One of the

assumptions

of Christaller'sclassical formulation

of central

place theory is that consumers

will always go to the closest market,

n order to

minimize he energy and time

costs of transportation.Crissman

12) has argued

that this will not always be true because . . . motivations for marketing at one

place rather hananotherare exceedingly

complex and varyfromone context to

another. Specifically, Crissmansuggests that . .

.

marketing

areas for par-

ticulargoods associated with

given levels in the central place

hierarchy, vary

significently n size dependingon the level in the hierarchyof

the center sup-

plying them (12, pp. 347, 393). Trips to higher-order enters

may be more

frequent than theoretically predicted,he argues, because of the

possibilityof

multipurpose trips, . . . price differentials between low and high-level

centers

owing to fewer middlemen at the former. . . , and lower prices

in larger centers

due to

. . .

competition between

multiple firms offering the same goods

.

.

Added to these economic considerations,he argues, are a variety of social,

political, religious, and recreationalreasons for going to

one town rather

han

another. We

might expect that the beauty and impressiveness

of

the massive

civic, ceremonial,and palace

constructions, manifesting he power

of the high-

est

administrativeevel, would

be an added ncentive to visit the

majorpolitical

center

in

spite

of

the

tribulations

of travel. For a variety of reasonsthen,

the

presence of a primatecentercould retard he growth of other centers

because

people

will

tend

to

go

out of

their way to complete transactions

n the

large

center, bypassingbetter-locatedplaces.

(b) Carol

Smith

(52)

has described a situation

in

highlandGuatemala

which

may add to

our

understanding

f the nature of primate settlement

distribu-

tions. There the massedpurchasingpower of the Ladinoelites,

who control

the

government

from a few centers, has resulted in a situation

in which

rural

production

s

oriented

to

these

largecenters alone. Thus all economic

networks

in

the region

. .

. converge on a single center rather than

on

different

nodes

of

a

multi-centered,unboundedsystem. Because of this convergence,

there

is

no

competitionamongequivalent

centers for the

commerce

of

smaller

centers

(52

p. 100).

The

outcome of these

processes,

she

argues,

will be

the

growth

of

solar

or

dendritic

market

ystemsinwhich secondarycenters

in

the

region

are poorly developed, and most production s gearedfor the primatecenter.

Such

situations lend themselves

to

poor economic development

of

the hinter-

land,

since

populations

ar

away

from

the maincenter cannotparticipate

n the

society's

economic

system as

effectively as those closer in (52;see also

Johnson

23).

8/12/2019 2949313 mercado

10/17

STUDIES

OF

CITIES

257

(c)

According to central

place

theory,

secondary

centers

should

arise far

away

from the

majorcenter,

where

the

pull of the

major

center is weakest due

to

the

friction of distance.

This is

not realizable

in

some

cases, however,

because

secondary centers far enough away from the major center to effectively compete

with it

may be near

the

borders of

the society.

Because

their

market

ranges are

therefore

distorted

and

smaller than

necessary

to support

higher level

functions,

they will

tend to

remain poorly

developed,

the

demand

being

met

for higher-

order goods

only

at the main

center.

According

to Johnson

(23, p.

131), border

.

. .

cities are inherently

fragile economically

since

national

borders

artificially

cut

up

geographically

complementary

regions

(see also 51,

p.

179).

(d)

Because

long-distance

exchange

is

often

directly

administered

or is at

least

under

some

degree

of political

control,

high-ranking

administrative

centers

are

often also the foci of interregional exchange activities, and this, too, could add to

the already

overbearing

domination of

a region

by a primate

center

(cf

Vap-

narsky

55).

Disembedded

Capitals

In cases

where

the highest-order

goods

in a society

have a range

which is

less

than

the extent of

the

society as

a whole,

there

should

be multiple

highest

ranking

economic

centers

instead

of the

primate pattern.

In

this

case,

I suggest,

it

is less

likely that

the

highest-order

place

in the administrative

hierarchy

will

be

located in one of the commercial central places, for at least two reasons:

(a)

The location of any

one of the

multiple

commercial

central places,

which

might be optimal

for

servicing

its region

but not

the

whole society,

might

not

be

suitable

for the location

of central administrative

transactions

pertaining

to

the

whole society. Suboptimal

locations

of highest

ranking

central

places

in

the

administrative system,

as

I

mentioned

above,

increase the

likelihood

that

noise

could enter

into communications

with

poorly serviced

zones,

de-

creasing

the

effectiveness

of the administration.

(b)

Merchants

in existing high-order

commercial

central

places

might

resist

placement

of

high-order

administrative

functions

in any

existing center

(except

their

own),

because

of the

commercial advantages

that

would accrue

to that

center, especially

insofar

as such

placement

would

increase

the

prestige

of

the

one center

and mass

the

purchasing

power

of

high

ranking

elites

there.

I

predict

that in systems

with

multiple

high-order

commercial

central

places

of

nearly equivalent

rank,

the

political

capital

will

be

located

in

a neutral

position,

away

from

existing

commercial central

places.

This situation would also

obtain

in

a region

inhabited

by

a

group

of autonomous

political

units

joined

in a

league

or confederacy, perhaps

for

the purpose

of

taking

advantage

of

their

mutual

military capabilities.

in

interregional

warfare

or control

of interregional

long

distance exchange, but where all of the co-joining units are of nearly equivalent

power.

There are,

in

other

words, situations

in which

one

would expect

the

highest-order

decision-making

institution

to be spatially

disembedded

from

the remainder of the

central-place

hierarchy.

There

appear to

be several

differ-

ent kinds of

disembedded

capitals.

8/12/2019 2949313 mercado

11/17

258

BLANTON

CAPITAL CENTERS The

capital

is a

permanent but

neutrally

located

special

function

community involved

primarily

in

decision

making at the

regional level.

Examples

of

this pattern

include

Washington DC (which was located

away from

the existing centers of Boston, New York, and Philadelphia, midway between

the

southern

and northern states, at the

insistence of

the southerner, Thomas

Jefferson),

pre-Periclean Athens,

Brazilia, Ottawa,

New Delhi, and Canberra. I

have

argued that early

Monte Alban, the ancient

Zapotec capital in

Oaxaca,

Mexico, was

a center of this type

(Blanton 9).

ROVING

PALACES

Highest-ranking elites move

around periodically or at ran-

dom

from center to

center, never staying long enough

in one place to

augment

permanently the influence

of any given

center. Examples of this pattern

include

the

roving palaces of

ancient Egypt

and certain periods and

regions

in

pre-

modern

Europe.

According to Russell (43), the failure

to establish

permanent

places for ruler and court

in Medieval

Germany and Iberia precluded

the rise of

single, dominant central

places.

TEMPORARY

CAPITALS

Rulers build entirely new

capitals

when

they

come

into

office. Part of

Coe's (11) argument that the

ancient inland Cambodian

capitals were not

true

cities was that each new ruler

built a new

facility,

an

impossible situation, Coe

thought, for a

commercial center. It is

likely that he

was

dealing instead

with

a

situation

in

which the

capitals

were

disembedded

from the

remainder of the central place

hierarchy.

Unfortunately, the paucity of

archaeological evidence

makes it

difficult to test this hypothesis.

I

feel it would be

worthwhile to

investigate certain of the dynamic

properties

of the

two

kinds of

central place

hierarchies just

described. Specifically,

it

occurs to me

that the first

pattern, the primate pattern

with a weakly

developed

hierarchy of commercial

central places,

might exhibit less long-term

stability

than the

second

pattern,

in

which the

whole population is serviced

by a

more

completely

developed

series of marketing central

places. The class-like

differ-

ences that

are likely

to

develop, given the

primate pattern, between

the well-

serviced,

relatively

affluent

groups

in

and near

the main

center

versus

those

in

poorly serviced, distant areas, could produce internal tensions that could erupt

as warfare or

revolution.

Too,

the

primate

pattern might

be a more brittle

one

because

populations

distant

from the main center

might tend to break

away

politically

and

economically

from

the

society as a whole, especially

if

allegiance

to an

adjacent

system

with

a

well-developed commercial

central place hierarchy

becomes

a

possibility.

Political

Control of

Marketing Hierarchies

Although

classical central

place theory predicts a lack of

spatial

isomorphism

between administrative and commercial central place hierarchies (with the

exception

of

the

primate pattern described

above),

in

reality

there is often

considerable

spatial intertwining

of the two

types

of

central

institutions.

One

reason for this

undoubtedly has

to do with the

energy

and

time

savings

derived

from the

agglomeration

of

functions.

Skinner writes

concerning

rural

China:

8/12/2019 2949313 mercado

12/17

STUDIES OF CITIES

259

That the market town is a natural site for administration f its dependent area is

obvious: Marketing periodically brings to town representatives of households

throughout he administrative nit and village leaders can, on marketdays, readily

consult their administrative uperiors 50, p. 382).

Another factor to consider in explaining the overlap of political and commercial

hierarchies is the extent to which the decision making central institution eyes the

exchange system as a source of revenue and power. In this case,

locating

administrative and marketing activities together facilitates control and taxation.

In spite of the obvious advantages of conformity between economic and de-

cision hierarchies, however, there still exists the built-in incompatability be-

tween the two. This is probably why Skinner found little overlap between

administration and market places in rural China on the whole, especially in the

lower levels of the central place hierarchy (48, p. 9).

I

suggest that the most fruitful method for explaining the cross-cultural and

diachronic variability in relations between administrative and marketing

hi-

erarchies will be to investigate the following three variables:

(a) The power of the state: This refers specifically to the extent to

which the

administrative institution is able to distort the economically desirable K=3 and

K=4 patterns into the administratively more efficient K=7 pattern. According

to

Smith (52, p. 99) the K=7 pattern usually resembles the solar marketing

pattern. Here,

. .

. poor market articulation occurs because

the

requirement

of

force overrides market efficiency. . A K=7 system almost always suggests

imposition of the economic system by rulers . . . I predict that such administra-

tively imposed patterns would always be temporary, because the more an

administration meddles

in

markets, the less efficient marketing becomes.

For

example,

the K=7

pattern could easily lend itself to monopolistic pricing,

since

equivalent centers

no

longer compete. Producers, under these conditions, might

choose

to

participate

less

in

the market

institution, lowering

over-all

productiv-

ity, and thereby lowering state revenues from the market. This would coincide

with the

added costs incurred by the administration in its attempts to gain control

over

markets. These two factors operating together should force

lessened

ad-

ministrative control of markets, probably first in lower ranking markets, which

are

the

most

numerous and therefore require the most personnel

to administer.

This

process is nowhere better illustrated than in Skinner's description

of the

attempts

of

the communist

regime

in

China

to

administer

peasant marketing

(50).

The

first such attempts resulted in considerable dislocations

in the

market

and

in

reduced

productivity.

The

response by

the

regime

to

the

lowered

produc-

tivity

was to

attempt

to

obviate rural marketing by absorbing

it .

. .

entirely

into

an expanded official structure, thereby shifting

onto hundreds

of

thousands

of

local

cooperative branches the functions of the standard

market

(50, p. 365).

This

tactic only exacerbated the problem, until finally the regime

reversed

its

position: The rightist solution [to the marketing problem], which by August,

1956,

had

carried the

day, was

to

overcome

the malfunctions

in the

existing

system by relaxing

controls

and

giving

freer rein to

the

marketing

mechanism.

(b) The relative power of the state and participants in the

market institutions:

A

politically powerful marketing institution might force the diminution

of

admin-

8/12/2019 2949313 mercado

13/17

260 BLANTON

istrative

distortions of the

marketing hierarchy. This process is well illustrated

in

the history of Medieval Europe,

where the emerging

commercial middle class

gradually

swamped the power of the nobility. To quote

Pirenne (40, p. 203), the

legal codes of the commercially oriented cities . . . not only did away with

personal

servitude and r-estrictionson land, but also caused

the disappearance of

the

seignioral rights and fiscal claims which interferred

with the activity of

commerce and

industry. This

resulted

in

the elimination of, among other

things,

market tolls, seigniorial monopolies on ovens and

mills, and the right of

shelter for the nobility.

The possibility exists that there

could develop in such

cases a kind of market-

state

dynamic, such that,

if

for

some reason the state begins to lose control over

the market institution, an

increase in market

productivity would result, in-

creasing in turn the power of

the marketing institution,

which could further

exacerbate the

power and revenue losses of the

administrative institution-a

deviation-amplifying,

mutual-causal process, in the phraseology of Maruyama

(33). Administrative expansion in

control of markets,

however, will normally be

dampened as a result of revenue

losses incurred in distortion of the marketing

hierarchy, in combination with

increased costs of

administration.

(c) The

information processing capacity of

administrative

institutions:

This

refers to

the extent to which a

decision making institution can absorb an increase

in

the amount of

information in the environment, for

example,

in

the context

of

an

expanding, evolving system.

Any institution has a finite transactional capac-

ity, limited by energy, personnel, and time, allowing it to respond only to a finite

communications load.

According to Meier:

Communications

oad is

best measuredas

the rate

of

requests

for service

or other

forms of

satisfying response,

such as

cogent explanations

as to

why

the service

cannotbe

provided.

It

represents he initiation atefor social transactions ffected

by

the

institution.The

output,

or

transaction ompletion,rate s

whatgenerates

rewards

for

the institutionover

the long run.When the outputrate fails to

keep up

with the

initiationrate, some queues and

backlogsdevelop. If all other

factors remainequal,

the

resultant

performance f the

institutionbeginsto deviate more

and

morefrom

the

idealas load

ncreases, untila peak n

outputrate s reached.As certainresources

and

internal tocks-on-hand reexpended, the outputbecomes less with increasingoad.

The

output

rate

may drop

precipitately resulting n breakdown]

r

it

may seek

a level

which is

'good enough to get by.' The

capacity of the organization

or

completing

transactionswill lie

somewhere

between the peak performance hat could

not be

maintained nd

the level chosen for 'satisficing'. . . In this

framework he capacity

of an

institution or

completinga flow of transactions s

equivalent

to the channel

capacity

of

a

communications

ystem

for

coding

and

decoding messages (34, pp.

79-80).

(See

also

J.

G. Miller

35

and

Wright 63.)

Several different kinds

of

institutional

responses

to

communications

overload

are possible, as Meier points out (including expansion in size if the institution

can

capture

more

energy

to fund

growth),

but

the

kind

of

response

most relevant

to

this

discussion

is for

the institution

to

redefine

its

limits inward.

An adminis-

trative

institution,

in the

context of stress due to communications

overload

(for

8/12/2019 2949313 mercado

14/17

STUDIES

OF CITIES 261

example in an

evolving system), might

respond by

allowing

certain

previously

administered transactions to operate

autonomously.

I

suggest that in

such cases

one of the

first areas

in

which control

hierarchies would

relinquish control

would

be in the context of exchange-related institutions. This is because economic

institutions will tend, in

the absence of

administration, to exhibit

self-regulation,

or

at least

will

tend

to

operate with a

minimum of administrative

regulation. This

response

would free energy and

personnel for more

urgent problems in the areas

of offense

and defense,

long-distance trade, and

adjudication.

The arguments

presented

above, based especially on

Skinner's analyses, strongly

suggest that

economic

institutions

actually do better with

minimal administration

because

of

the built-in

incompatabilities of the

marketing and

administrative central place

hierarchies.

This kind of response to

information

overload could explain the

transition from

redistributive economies (the

most highly

administered ex-

change

systems)

to

markets, as well as being a

possible explanation for

why

an

administration might

relinquish control

of existing markets.

CONCLUSIONS

I

have argued that

traditional theories of cities

utilized by

anthropologists fail

because

they

cannot

adequately account

for

the cross-cultural

variability

in

the

nature

of

cities,

and

because they are

inadequate for dealing with

the dynamic

properties of cities and

the societies of

which they are a part. I

have suggested

that

one

avenue anthropologists could

take

to

resolve the

shortage

of

adequate

theory

is

to

adopt

a

regional

approach

to

studies

of

cities,

based

on

the

powerful

central

place theory

of

geography.

This

theory focuses attention

on the

dis-

position

of central

institutions over the

landscape. Those

places

that are the

foci

of

central

institution

transactions

are a society's

system

of

cities

and towns-its

central-place hierarchy. This

theory not

only provides the means for

explaining

some

of the

cross-cultural

variability

in

systems

of

cities

and

towns,

but also

provides

a

means for

dealing with certain of the

dynamic properties of societies.

To

accomplish

this,

I

argued, the

regionally oriented

researcher should

investi-

gate the spatial relationships of the two basic kinds of central institutions, those

that are economic and

those that

pertain to information processing

and decision

making.

I

outlined the

basic kinds

of

articulation

between

these

two

genre

of

central

institutions,

described

the

patterns

in

the central

place

hierarchy

which

result

from

these

relationships, and explored some of

the implications

of

these

varying patterns for

understanding

social change. The three patterns

explored

are the

following:

(a)

The

primate pattern

obtains

in

a

region

in

which the

major

commercial

central

place

is also

the

political capital.

This

center

will

tend

to

dominate the

region

in

such a way that the

development of

secondary centers,

as

predicted by

Christaller's economic location principles, is retarded. This can result in a

dendritic

market pattern on

the

peripheries, such that regions distant from the

main

center are poorly

serviced

economically. This pattern is

typically

associ-

ated with

the dual

economies

described

by Johnson

(23),

in

which

great

8/12/2019 2949313 mercado

15/17

262

BLANTON

wealth differentials develop

between those populations close

to or

in

the

center,

and those

populations

in

distant, poorly serviced

areas.

Such societies, I

argued,

might exhibit relatively little long-term stability because these

wealth

differ-

entials could lead to internal conflict, and because poorly serviced peripheral

regions might tend to break away

politically and economically from the society

as a whole.

(b) The second pattern I

described is one in which there is more complete

development of the commercial central place hierarchy, as predicted by Chris-

taller. The primate pattern is absent, in part, because the political capital is

typically disembedded from

the marketing hierarchy, and so does not

artificially distort that hierarchy away from what would be expected on the-

oretical grounds alone (i.e.

following Christaller's K=3 and K=4 locational

principles).

(c) Lastly,

I

discussed

the

nature

of

the lower ends

of

central place

hi-

erarchies. The nature and spatial

distribution of low-order central places in a

society,

I

argued, depends largely

on the extent to which a decision making

institution is able to distort the economically desirable K=3 and K=4 pat-

terns

into

the administratively

desirable K=7 pattern, in order to more closely

regulate and tax rural production.

Following the work of Skinner (50) as an

example,

it is

clear that marketing

efficiency depends on maintenance of the

K=3

or K=4

patterns.

Administrative meddling

in

the marketing hierarchy

brings

a

reduction

in

marketing

efficiency, and therefore, overall declines

in

productivity of participants in the marketing institution. As a result, I suggest,

administrative

distortions of

commercial central place hierarchies

should tend

to

be short-term, since an administration loses revenues at the same time that

it is

incurring the added costs of market administration.

I

hope this brief, and not nearly

complete, exercise in central place theorizing

will

be an inducement to other

anthropologists interested in cities to adopt a

regional approach as one method

for explaining the variability and dynamic

features

of

systems

of

cities and towns.

Relatively few anthropologists

have

employed

such

an

approach.

The

prominent works are

those

I

have

already

described at length, those of Skinner (48-50), Smith (52), and Crissman (12).

To

this list I can add the publications of Johnson (24-26), Wright (63), Adams (1),

and

Adams & Nissen (2), for ancient

Greater Mesopotamia; Marcus (31),

and

Hammond

(18)

for

the ancient

lowland Maya area; the publications

of

par-

ticipants in the Valley of Mexico

Project (Blanton 8, Millon 36, Parsons 39,

Sanders

45,

Wolf

62);

as well as

Blanton (9)

and

Kowalewski (27), writing

about

the ancient Valley of Oaxaca. This small but growing movement,

I

feel,

will

prove

to

be

a

significant new

addition

to

anthropological

studies

of cities.

ACKNOWLE

DG

MENTS

I

am grateful for helpful comments

received on an earlier version

of this

paper

from Gary Feinman, Greg Johnson, Steve Kowalewski, Susan Lees, John

Pfeiffer, John Speth, Donna Taylor,

and

Paul

Wheatley. Responsibility

for

any

errors

is

my

own.

8/12/2019 2949313 mercado

16/17

STUDIES

OF

CITIES

263

Literature

Cited

1.

Adams, R.

McC.

1965.

Land

Be-

hind

Baghdad:

A

Historv

of Set-

tlement on the Diyala Plains. Univ.

Chicago Press

2.

Adams, R.

McC.,

Nissen,

H.

J.

1972.

The

Uruk

Countryside.

Univ.

Chicago

Press

3.

Arensberg,

C.

M.

1968.

The

urban

in

crosscultural

perspective.

In

Ur-

ban

Anthropology:

Research

Per-

spectives and

Strategies,

ed.

E.

Eddy,

2:3-15.

Proc.

South. An-

thropol. Soc.

Athens:

Univ.

Geor-

gia

Press

4.

Berry, B. J.

L.

1961.

City size dis-

tribution

and

economic

develop-

ment. Econ. Dev. Cult. Change

9:573-88

5.

Berry, B.

J. L.

1967.

Geography

of

Market

Centers and

Retail

Distribui-

tion.

Englewood

Cliffs:

Prentice-

Hall

6.

Berry,

B.

J. L.,

Garrison,

W.

L.

1958.

Recent

developments in

cen-

tral-place

theory. Proc.

Reg. Sci.

Assoc.

9: 107-20

7.

Berry,

B.

J.

L.,

Pred, A.

1961.

Ceni-

tral

Place

Studies: A

Bibliographv

of

Theory and

Applications.

Philadelphia:Reg. Sci. Res. Inst.8.

Blanton,

R. E.

1972.

Prehispanic

Settlement

Patterns

of the

Ixta-

palapa

Peninsula

Region, Mexico.

University

Park:

Penn.

State

Univ.

Dep.

Anthropol.

9.

Blanton,

R.

E.

1976.

The

origins of

Monte

Alban. In

Cultural

Chainge

and

Continuity,

ed.

C.

Cleland.

New

York:

Academic

10.

Christaller,

W.

1933.

Die

Zentralen

Orte in

Suddeutschland.

Transl.

C. W.

Baskin, 1966.

Central

Places

in

Southern

Germany.

Englewood

Cliffs:Prentice-Hall

11.

Coe, M. D.

1961.

Social

typology

and

tropical

forest

civilizations.

Comp. Stud. Soc.

Hist.

4:65-85

12.

Crissman,

L.

W.

1973.

Town

and

country: Central

place

theory

and

Chinese

marketing,

with

particular

reference to

southwestern

Chan-

ghua

Hsien,

Taiiwan. PhD

thesis.

Cornell

Univ.,

Ithaca,

NY

13.

Epstein, A. L.

1967.

Urbanization

and

social

change

in

Africa.

Curr.

Anthropol.

48:275-95

14.

Flannery,

K.

V. 1972.

The

cultural

evolution of civilizations. Ann. Rev.

Ecol.

Syst.

3:399-426

15.

Friedmann,J. 1961.

Cities

in

social

transformation.

Comp.

Stud.

Soc.

Hist.

4:86-103

16. Gunawardena,

K. 1964. Service

centers

in southern Cevlon.

PhD.

thesis. Univ. Cambridge,England

17. Haggett,

P. 1965. Locational

Analy-

sis in Humnan

Geographly.

New

York: St. Martin'sPress

18. Hammond,

N.

1972.

Locational

models and the site of

Lubaanturn:

A classic Maya centre. In Models

in

Archaeology,

ed.

D.

Clarke,

757-800. London: Methuen

19. Hauser, P. M. 1965.

Urbanization,

an overview.

In

The Study

of Ur-

banization,

ed.

P. M.

Hauser,

L.

Schnore, p.

12. New York:

Wiley

20. Hauser, P. M. 1965. The

folk-urban

ideal types: Observationson the ur-

ban-folk

and

urban-rural

di-

chotomies

as forms of Western eth-

nocentrism. See Ref. 19,

503-17

21. Hodder, I.,

Hassall

M. 1971.

The

non-random spacing

of Romano-

British walled towns.

Main

6:391-407

22. Jacobs, J. 1969. The

Economv

of

Cities.

New

York: Random

House

23. Johnson,

E. A. J. 1970. The Organi-

zation of Space

in

Developing

Countries. Cambridge:

Harvard

Univ. Press

24.

Johnson, G.

A.

1972.

A

test of the

utility of central place

theory

in

ar-

chaeology.

In

Man, Settlement,

and

Urbanism,

ed.

P. J.

Ucko,

R. Tri-

ngham, G.

W.

Dimbleby, 769-85.

London: Duckworth

25. Johnson, G.

A. 1973.

Local Ex-

change

and

Early

State

Develop-

ment

in

Southwestern

Iran.

An-

thropol.

Pap. 51.

Ann

Arbor: Univ.

MichiganMus. Anthropol.

26. Johnson, G.

A.

1975.

Locational

analysis

and

Uruk local exchange

systems. In Ancient Civilization and

Trade,

ed. J. A.

Sabloff,

C. C.

Lamberg-Karlovsky, 85-339.

Albu-

querque:Univ. New

Mexico Press

27.

Kowalewski, S.

1976.

Prehistoric

settlement patterns of

the central

part of the valley of Oaxaca,

Mex-

ico. PhD. thesis. Univ. Arizona,

Tucson,

Ariz.

28.

Lewis,

0.

1951.

Life

In

a

Mexican

Village. Tepoztlin

Restudied. Ur-

bana: Univ.

Illinois

Press

29. Lewis,

0. 1965. The folk-urban

ideal types:

Further

observations

on

the folk-urban continuum and ur-

banization with

special

reference

to

Mexico

City.

See Ref.

19,

491-503

30.

Losch,

A. 1940.

Die riumliche Ord-

nung der Wirtschaft.

Transl.

H.

8/12/2019 2949313 mercado

17/17

264 BLANTON

Woglom,

W.

F.

Stolper,

1954.

The

Economics

of Location.

New

Ha-

ven:

Yale Univ.

Press

31.

Marcus, J.

1973.

Territorial

organi-

zation of the Lowland Classic

Maya.

Science

180:911-16

32. Martindale,

D. 1958.

Prefatory

re-

marks:

The

theory

of

the city.

In

The

City, M.

Weber,

9-62.

Transl.

D. Martindale,

G. Neuwirth,

New

York: Free

Press

33.

Maruyama,

M. 1963.

The

second

cybernetics:

Deviation-amplifying

mutual-causal

processes.

Am.

Sci.

51:164-79

34. Meier, R.

1962.

A Communications

Theory

of Urban

Growth.

Cam-

bridge:M. I. T. Press

35.

Miller, J.

G.

1965.

Living systems:

Basic concepts.

Behav.

Sci.

10:193-257

36.

Millon,

R.

1973.

Urbanization

at

Teotihuacin,

Vol. I.

The

Map.

Aus-

tin:

Univ.

Texas

Press

37.

Moore,

K. 1975.

The city

as con-

text:

Context

as process.

Urban

An-

thropol.

4:17-25

38.

Park,

R. 1925.

The

city: Sug-

gestions

for

the

investigation

of

hu-

man behavior

in

the

urban

environ-

ment.

In The City,

ed.

R. Park,

E.

Burgess, R. McKenzie, pp. 1-46.

Univ. Chicago

Press

39.

Parsons, J.

R. 1972.

Prehistoric

Set-

tlement

Patterns

in the

Texcoco

Re-

gion,

Mexico.

Mem.

3, Mus.

An-

thropol.,

Univ. Michigan.

447

pp.

40. Pirenne,

H. 1925.

Medieval Cities:

Their

Origins

and the

Revival

of

Trade, p.

203.

Princeton Univ.

Press

41. Redfield,

R. 1941.

The Folk Cul-

tures

of

Yucatan. Univ.

Chicago

Press

42.

Redfield,

R., Singer,

M. 1954.

The

cultural role of cities. Econ. Dev.

Cult. Change

3:53-73

43.

Russell, J. C.

1972.

Medieval

Re-

gions

and

Their Cities,

p.

36.

Bloomington:

ndianaUniv. Press

44.

Sahlins,

M. D. 1968. Tribesmen.

Englewood

Cliffs:

Prentice-Hall

45.

Sanders,

W.

T. 1965. Cultural Eco-

logy of

the

Teotihuacin

Valley,

Mexico. University

Park:

Penn.

State

Univ.

Dep.

Anthropol.

46.

Service,

E.

R. 1971.

Primitive So-

cial

Organization.

New York:

Ran-

dom House.

2nd ed.

47.

Sjoberg,

G.

1960. The Preindustrial

City,

Past

and Present. Glencoe:

Free Press

48. Skinner, G. W.

1964.

Marketing

and social

structure

in rural China:

Part I. J. Asian Stud. 24:3-43

49. Ibid

1965. Marketing and social

structure

in rural

China: Part

II,

195-228

50.

Ibid.

Marketingand social structure

in ruralChina:Part

III, 363-99

51.

Smith,

C.

A.

1974.

Economics of

marketing

systems: Models from

economic

geography.Ann. Rev.

An-

thropol.

3:167-201

52.

Smith,

C. A. 1975. Examining

stratificationsystems

through peas-

ant

marketing

arrangements:An ap-

plicationof some models from eco-

nomic

geography. Man 10:95-122

53. Stafford,

H. A. Jr. 1963. The func-

tional

bases of small towns. Econ.

Geogr.

39:165-75

54.

Taylor,

D.

1975.

Some locational

aspects

of middle-range

hierarchical

societies.

PhD. thesis. City Univ.

New

York, NY

55.

Vapnarsky,

C. 1969. On rank-size

distributionsof cities:

an ecological

approach.

Econ. Dev. Cult. Change,

17:584-95

56.

Wheatley, P. 1963.

What

the

greatness of a city is said to be:

Reflections

on Sjoberg's

Pre-

industrial

City.

Pac. Viewpoint

4:163-88

57. Wheatley, P. 1971.

Pivot of

the

Four

Quarters, A Preliminary

In-

quiry

into the Origins and Character

of

the

Ancient Chinese City,

Chicago:Aldine

58.

Wheatley,

P. 1972. The concept

of

urbanism.

See

Ref. 24, 601-37

59.

Wirth,

L.

1925.

A

bibliography

of

the urban community.

See Ref.

38,

p.

172

60. Wirth,L. 1938.Urbanismas a way

of life.

Am.

J.

Sociol.

44:1-24

61. Wolf,

E.

R. 1966.

Peasants, p.

11.

Englewood

Cliffs:

Prentice-Hall.

116

pp.

62.

Wolf,

E.

R.,

ed.

1976.

The

Valley

of

Mexico:

Studies

in

Ecology

and

So-

ciety. Albuquerque:

Univ. New

Mexico

Press

63.

Wright,

H. T.

1969.

The

Adminis-

tration

of

Rural

Production

in

an

Early

Mesopotamian

Town.

An-

thropol.

Pap.

38.

Ann Arbor: Univ.

Michigan Mus. Anthropol.